The goal of this research is to identify etiologic factors relating to two surgical outcomes: death and failure-to-rescue, in order to improve the credibility and usefulness of surgical outcome measurement. The project will use Medicare claims data, merged with Pennsylvania claims data including admission severity score, to construct a multivariate matched case-control study of surgical outcomes. Detailed chart review of randomly selected cases and controls will constitute the study population of 4150 patients, and will enable the construction of conditional logistic regression models to answer questions regarding the influence of in- hospital factors (caregiver and caregiver actions) and across-hospital factors (hospital characteristics) on surgical outcomes. The case control analysis will be supplement with data from a detailed hospital survey, which also will be used to study hospital and caregiver characteristics related to outcome in a larger claims data analysis. Specific Aims are as follows: (1) To determine, for general surgical and orthopedic inpatient procedures, those hospital and caregiver factors (structure and process factors that can potentially be changed to improve quality of care) most closely associated with outcomes (death or failure-to-rescue), adjusting for patient factors; (2) To determine important etiologic factors (structure and process variables) leading to better or worse outcomes of death and failure-to-rescue based on an in-depth analysis of charts using a case-control methodology to allow for improved power. By appropriate matching, caregiver effects may be separated from hospital effects; (3) To determine the correlation between hospital rankings based on death rate, complication rate, and failure rate and those hospital characteristics found to be important from the case control study. Determination of etiologic factors related to outcome will improve the credibility of outcome measurement and aid in the appropriate use and understanding of such measures.